In this assignment I explore different strategies used in building and training a Language Translator. I use Seq2Seq learning to convert sequences from English to Hebrew. I also include techniques such as bi-directional learning and Attention mechanism which serve as the building blocks of advanced transformer-based NLP models such as GPT, Llama etc. .

I initially wanted to build a model to translate English to Tamil which happens to be my Native Language so that I could work with the translations easily. But due to less number of available sentence pairs (207) on the https://www.manythings.org/anki/ website for this language (the website used in my class) , I picked Hebrew.

Though Hebrew has been a language of fascination for me for a while now, there are a few other important reasons I picked Hebrew for learning machine translation.

• Hebrew is written from right to left, it will be interesting to see if Bi-LSTM and Attention produces better results.

• English and Hebrew are from completely different language-families and roots. (https://webspace.ship.edu/cgboer/languagefamilies.html)
  They are from completely different regions and time-periods and have isolated places of origins.

  • English -> Region: West Germany, Language family: Indo-European, Root: Germanic
  • Hebrew -> Region: Israel, Language family: Afro-Asiatic, Root: Semitic

• They share almost a near-zero lexical similarity (https://en.wikipedia.org/wiki/Lexical_similarity) and less genetic relationship and linguistic interference. (https://en.wikipedia.org/wiki/Genetic_relationship_(linguistics))

• English-Hebrew also has a good number of sentence pairs available on the https://www.manythings.org/anki/ website(127856)

Image Source: https://alternativetransport.wordpress.com/2015/05/05/34/

• Used Hebrew Tokenizer (github.com/YontiLevin/Hebrew-Tokenizer) to parse and clean Hebrew text (Canonical normalization used for English does not work)
• Converted text to sequences (reversed sequence list for Hebrew)
• Models : Seq2Seq, Bi-directional Seq2Seq (Bi-LSTM), Seq2Seq with Attention layer
• Tuning: Latent dimensions, training epochs, dropout, activation functions
• Sampling: Greedy and Multinomial (with different temperatures)

Link to Notebook

This repository is Software Carpentry and Data Carpentry‘s template for creating websites for instructor training workshops.

1. Please do not fork this repository directly on GitHub. Instead, please use GitHub’s importer following the instructions below to create a website repository for your workshop.

2. Please do your work in your repository’s gh-pages branch, since that is what is automatically published as a website by GitHub.

3. Once you are done, please send your repository’s URL to the Software Carpentry administrator. We build the list of workshops on our websites from the data included in your index.html page. We can only do that if you customize that page correctly and send us a link to your workshop website.

If you run into problems, or have ideas about how to make this process simpler, please get in touch. The pages on customizing your website, the FAQ, and the design notes have more detail on what we do and why. And please note: if you are teaching Git, please create a separate repository for your learners to practice in.

1. Log in to GitHub. (If you do not have an account, you can quickly create one for free.) You must be logged in for the remaining steps to work.

2. Go to GitHub’s importer.

3. Paste the url of this repo as the old repository to clone: https://github.com/carpentries/training-template.

4. Select the owner for your new repository. (This will probably be you, but may instead be an organization you belong to.)

5. Choose a name for your workshop website repository. This name should have the form YYYY-MM-DD-site-ttt, e.g., 2016-12-01-miskatonic-ttt, where YYYY-MM-DD is the start date of the workshop.    for online workshops, choose online as site

6. Make sure the repository is public.

7. At this point, you should have a page like this:

   

   You can now click “Begin Import”. When the process is done, you will receive a message like “Importing complete! Your new repository gvwilson/2016-12-01-miskatonic is ready.” and you can go to the new repository by clicking on the name.

Note: some people have had intermittent errors during the import process, possibly because of the network timing out. If you experience a problem, please re-try; if the problem persists, please get in touch.

1. Go into your newly-created repository, which will be at https://github.com/your_username/YYYY-MM-DD-site-ttt. For example, if your username is gvwilson, the repository’s URL will be https://github.com/gvwilson/2016-12-01-miskatonic-ttt.

2. Edit the header of index.html to customize the list of instructors, workshop venue, etc. You can do this in the browser by clicking on it in the file view on GitHub and then selecting the pencil icon in the menu bar:

   

   Editing hints are embedded in index.html, and full instructions are in the customization instructions.

3. Alternatively, if you are already familiar with Git, you can clone the repository to your desktop, edit index.html there, and push your changes back to the repository.

   git clone -b gh-pages https://github.com/your_username/YYYY-MM-DD-site-ttt
   

   You should specify -b gh-pages because the imported repository doesn’t have a master branch.

   In order to view your changes once you are done editing, you must push to your GitHub repository:

   git push origin gh-pages
   

   Note: please do all of your work in your repository’s gh-pages branch, since GitHub automatically publishes that as a website.

4. When you are done editing, go to the GitHub Pages URL for your workshop and preview your changes. In the example above, this is https://gvwilson.github.io/2016-12-01-miskatonic.

Note: this template includes some files and directories that most workshops do not need, but which provide a standard place to put extra content if desired. See the design notes for more information about these.

Further instructions are available in the customization instructions. This FAQ includes a few extra tips (additions are always welcome) and these notes on the background and design of this template may help as well.

If you want to preview your changes on your own machine before publishing them on GitHub, you can do so as described below.

1. Install the software described below. This may require some work, so feel free to preview by pushing to the website.

2. Run the command:

   $ jekyll serve
   

   and go to http://0.0.0.0:4000 to preview your site. You can also run this command by typing make serve (assuming you have Make installed).

3. Run the command python bin/workshop_check.py index.html to check for a few common errors in your workshop’s home page. (You must have Python 3 installed to do this.) If you have Make installed, you can also run this command by typing make workshop-check.

In rare cases, you may want to add extra pages to your workshop website. You can do this by putting either Markdown or HTML pages in the website’s root directory and styling them according to the instructions give in the lesson template. If you do this, you must also edit _config.yml to set these three values:

1. carpentry is either “dc” (for Data Carpentry) or “swc” (for Software Carpentry). This determines which logos are loaded.

2. title is the title of your workshop (typically the venue and date).

3. email is the contact email address for your workshop, e.g., gvwilson@miskatonic.edu.

Note: carpentry and emailduplicate information that's inindex.html`, but there is no way to avoid this without requiring people to edit both files in the usual case where no extra pages are created.

If you want to set up Jekyll so that you can preview changes on your own machine before pushing them to GitHub, you must install the software described below. (Note: Julian Thilo has written instructions for installing Jekyll on Windows.)

1. Ruby. This is included with Linux and Mac OS X; the simplest option on Windows is to use RubyInstaller. You can test your installation by running ruby --version. For more information, see the Ruby installation guidelines.

2. RubyGems (the package manager for Ruby). You can test your installation by running gem --version.

3. Jekyll. You can install this by running gem install jekyll.

You can check the formatting of your header by running bin/workshop_check.py (which is invoked by make workshop-check). You must have Python 3 installed in order to do this, and you will also need the PyYAML module.

If you are teaching Git, you should create a separate repository for learners to use in that lesson. You should not have them use the workshop website repository because:

• your workshop website repository contains many files that most learners don’t need to see during the lesson, and

• you probably don’t want to accidentally merge a damaging pull request from a novice Git user into your workshop’s website while you are using it to teach.

You can call this repository whatever you like, and add whatever content you need to it.

We are committed to offering a pleasant setup experience for our learners and organizers. If you find bugs in our instructions, or would like to suggest improvements, please file an issue or mail us.

Features:

• Let you use slang-shaders with Capcom’s Mega Man X Legacy Collection.
• Fixes scaling artifact due to nearest-neighbour upscaling.

Download from here.

Using i686-w64-mingw32-gcc (cross compiling should work too):

# Download source
git clone https://github.com/xzn/d3d10-mmxlc.git
cd d3d10-mmxlc
git submodule update --init --recursive

# Create symlinks and patch files
make prep

# Build the dll
make -j$(nproc) dll

Some options to pass to make

# disable optimizations and prevents stripping
make o3=0 dll

# disable lto (keep -O3)
make lto=0 dll

Copy dinput8.dll, interp-mod.ini, and the slang-shaders\ directory to your game folders, e.g.:

• SteamLibrary\steamapps\common\Mega Man X Legacy Collection
• SteamLibrary\steamapps\common\Mega Man X Legacy Collection 2

interp-mod.ini contains options to configure the mod.

; Log API calls to interp-mod.log,
; [logging]
; enabled=true
; hotkey_toggle=VK_CONTROL+O
; hotkey_frame=VK_CONTROL+P

; Change interpolation mode and set up custom slang shaders.
[graphics]
; Use linear instead of point upscaling for the 2D games.
interp=true
; (WIP) Use linear scaling when possible for the 3D games.
; linear=true
; When using Type 1 filter, interp=true, and slang_shader* is not set,
; apply Type 1 filter over and over until it reaches screen size.
; enhanced=true
; Custom shader for X1~X6, needs Type 1 filter set in-game.
; slang_shader=slang-shaders/xbrz/xbr-lv2.slangp
slang_shader_snes=slang-shaders/crt/crt-lottes-fast.slangp
slang_shader_psone=slang-shaders/xbrz/xbrz-freescale-multipass.slangp
; Custom shader for X7~X8.
slang_shader_3d=slang-shaders/anti-aliasing/smaa.slangp
; (TODO) Custom render resolution for X7~X8
; render_3d_width=
; render_3d_height=
; Custom display resolution, e.g. 4K and so-on,
; Should be 16:9 as the mod currently does not correct for aspect ratio.
display_width=
display_height=

If all goes well you should now be able to start the game and see the overlay on top-left of the screen showing the status of the mod.

interp-mod.ini can be edited and have its options applied while the game is running.

Source code for this mod, without its dependencies, is available under MIT. Dependencies such as RetroArch are released under GPL.

• RetroArch is needed only for slang_shader support.
• SPIRV-Cross and glslang are used for slang_shader support.
• HLSLcc is used for debugging.

Other dependencies are more or less required:

• minhook is used for intercepting calls to d3d10.dll.
• imgui is used for overlay display.
• smhasher is technically optional. Currently used for identifying the built-in Type 1 filter shader.

Smooth scrolling bottom refresh control for table view with some objects in array.

Just add some dummy data into array with custom referesh control cell in table view or handle response from web service and
populate the data into the array with smooth animation from table view top with this two methods

Make sure these given below methods are very important to show the cell at the end of the table view with your conditional array objects
so that you can show the load more cell in table view with concerned API call for loading more data to an array.

Please go through the demo once and let me know if any query regarding this demo of bottom referesh control.

These were the methods to focus on while loading more data into cell with load more custom cell in table view as below:

Step 1 :

    - (NSInteger)tableView:(UITableView *)tableView numberOfRowsInSection:(NSInteger)section {
        return self.objects.count + 1;
    }

Step 2 :

    - (UITableViewCell *)tableView:(UITableView *)tableView cellForRowAtIndexPath:(NSIndexPath *)indexPath {
        static NSString *cellStandardIdentifier = @"Cell";
        static NSString *cellLoadingIdentifier = @"Loading";

        if (indexPath.row < self.objects.count) {
            UITableViewCell *cell = [tableView dequeueReusableCellWithIdentifier:cellStandardIdentifier forIndexPath:indexPath];
            cell.textLabel.text = self.objects[indexPath.row];
            return cell;
        } else {
            LoadingTableViewCell *cell = [tableView dequeueReusableCellWithIdentifier:cellLoadingIdentifier forIndexPath:indexPath];
            [cell.activityIndicatorView startAnimating];
            [self fetchMoreData];
            return cell;
        }
    }

Step 3 :

    - (void)fetchMoreData {
        static BOOL fetchInProgress = FALSE;

        if (fetchInProgress)
            return;

        typeof(self) __weak weakSelf = self;

        fetchInProgress = TRUE;

        // this simulates a background fetch; I'm just going to delay for a second
        dispatch_after(dispatch_time(DISPATCH_TIME_NOW, (int64_t)(1.0 * NSEC_PER_SEC)), dispatch_get_main_queue(), ^{
            typeof(self) strongSelf = weakSelf;
            if (strongSelf) {
                NSArray *indexPaths = [strongSelf addSomeObjects];
                [strongSelf.tableView beginUpdates];
                [strongSelf.tableView insertRowsAtIndexPaths:indexPaths withRowAnimation:UITableViewRowAnimationTop];
                fetchInProgress = FALSE;
                [strongSelf.tableView endUpdates];
            }
        });
    }

Step 4 :

    - (NSArray *)addSomeObjects {
        NSNumberFormatter *formatter = [[NSNumberFormatter alloc] init];
        formatter.numberStyle = NSNumberFormatterSpellOutStyle;
        NSMutableArray *indexPaths = [NSMutableArray array];
        for (NSInteger i = 0; i < 20; i++) {
            [self.objects addObject:[formatter stringFromNumber:@([self.objects count] + 1)]];
            [indexPaths addObject:[NSIndexPath indexPathForRow:self.objects.count - 1 inSection:0]];
        }
        return indexPaths;
    }